Welded structure fatigue life evaluation method based on substructure transient structural stress

Abstract

The fatigue life assessment method for welded structures based on transient structural stress in substructures comprises the following steps: Step 1: Constructing a reduced substructure model, a dynamic substructure model, and a weld substructure model; Step 2: Rigid-flexible coupling dynamic simulation; Step 3: Performing transient analysis using the weld substructure model; Step 4: Evaluating the fatigue life of the welded structure using the equivalent structural stress of the transient response. Compared to traditional transient analysis methods, this method considers both the rigid-flexible coupling dynamic simulation of the welded structure and the vibration characteristics of the welded structure itself, while significantly improving the efficiency of transient simulation analysis. It has important theoretical and practical significance for the fatigue resistance design of welded structures.

Description

Technical Field

[0001] This invention relates to the fields of mechanical and engineering technology, and more particularly to the field of performance testing technology for welded structures. Background Technology

[0002] Welding is one of the most important connection methods used in mechanical and engineering structures. Welded structures are often critical components; for example, most key components of rail vehicles, such as the car body, bogies, and suspension systems, are typically welded structures. The fatigue resistance of welds in welded structures is lower than that of the base metal. Therefore, the fatigue resistance of welds, especially those in critical load-bearing components, directly determines the service life of the welded structure.

[0003] Transient simulation analysis, compared to frequency domain analysis, more closely reflects reality and better reflects the dynamic characteristics of structures, such as unsteady-state and nonlinear properties, and has been applied to fatigue life assessment of welded structures. However, traditional transient analysis methods are difficult to widely apply due to the large size of the finite element model of the welded structure and the high computational time required. For example, a typical transient analysis of a rail-welded vehicle under one working condition takes approximately 10 hours. When comparing more than a dozen schemes, the exponentially increased and excessively high time cost causes designers to become overwhelmed, making it difficult to complete the design task. Therefore, there is an urgent need to research a new and efficient transient analysis method for fatigue life assessment of welded structures to solve practical engineering problems.

[0004] The substructure method is a comprehensive approach that divides the entire structure into several interconnected substructures and connects them together according to the coordination equations of displacement and force. Its model reduction theory provides the possibility of reducing transient simulation time. References: (1) Zheng Ling, Su Jintao, Wan Xinming, et al. A review of research on vibration substructure synthesis and correction methods [J]. Journal of Mechanical Engineering, 2019, 55(23): 120-135. (2) Zhang Caixia, Wan Chaoyan, Xie Suming. Finite element analysis of freight car body based on substructure technology [J]. Railway Vehicles, 2009, 47(01): 11-14+48. (3) DONDERS S, PLUYMERS B, RAGNARSSON P, et al. The wave-based substructuring approach for the efficient description of interface dynamics in substructuring [J]. Journal of Sound and Vibration, 2010, 329(8): 1062-1080. The above references are some studies and applications of the substructure method, but these references have not carried out research on weld fatigue life assessment based on substructure and reducing simulation time.

[0005] References: (4) Wang Qiushi et al. Fatigue life assessment method of bogieframe with time-domain extrapolation for dynamic stress based on extreme value theory[J].Mechanical Systems and Signal Processing,2021,159. (5) Fang Ji, Yang Chenxi, Han Zhengyan. Vibration fatigue analysis of SQ6 car body structure based on rigid-flexible coupling[J].Journal of Railway Science and Engineering,2019,16(08):2070-2076. (6) Tang Liming, Yang Min. Prediction of life of key welds of traction motor base based on frequency domain structural stress method[J]. Explosion-proof Motor,2021,56(04):17-20. The above references record that the traditional methods for assessing weld life by structural stress include quasi-static method, modal structural stress method and frequency domain method. The quasi-static method is to complete the dynamic simulation, count the rain flow of the load, and then accumulate the damage caused by the linear response of the weld equivalent to the structural stress according to the linear characteristics of the structure. It is evident that this method cannot consider the dynamic vibration characteristics of the structure itself. The modal structural stress method, limited by effective mass mode truncation and time-domain sampling frequency, and the frequency-domain structural stress method, have the limitation of only being able to analyze steady-state random vibrations.

[0006] References: (7) Liu Degang, Li Yabo. Simulation analysis of vibration fatigue of welded structure based on master SN curve [J]. Railway Vehicles, 2016, 54(01): 1-3+11+7. The above references record that structural stress transient analysis can reflect the vibration characteristics of the structure, but its simulation is very time-consuming. In addition, in order to simulate the dynamic characteristics of the vehicle system in the line operation, common rigid-flexible coupling dynamic simulation methods need to be considered. Summary of the Invention

[0007] To address the aforementioned problems in traditional fatigue life assessment methods for welded structures, this invention provides a fatigue life assessment method for welded structures based on transient structural stress in substructures.

[0008] The technical solution adopted by this invention to achieve the above objectives is: a method for evaluating the fatigue life of welded structures based on transient structural stress of substructures, the steps of which are as follows:

[0009] Step 1: Construct the reduced substructure model, dynamic substructure model, and weld substructure model;

[0010] Step 2: Dynamic simulation of rigid-flexible coupling;

[0011] Step 3: Perform transient analysis using the weld substructure model;

[0012] Step 4: Use the transient response of the welded structure to evaluate the equivalent structural stress fatigue life.

[0013] Step 1 is as follows:

[0014] 1-1: The geometric model of the welded structure from the previous step is meshed using the finite element method, with the mesh consisting of 2D or 3D elements.

[0015] 1-2: Construct a reduced substructure model.

[0016] The dynamic substructure method can transform a high-order linear system of equations into multiple low-order systems of equations, which can then be solved piecewise. Among these methods, Guyan's reduction theory is the most widely used.

[0017] Choosing the interface degree of freedom as the principal coordinate, denoted by B; and the internal degrees of freedom to be reduced as the secondary coordinate, denoted by I, then for a multi-degree-of-freedom system neglecting damping, its equation of motion during free vibration is:

[0018]

[0019] in, For the set of displacements of degrees of freedom, The set of displacements for the principal degrees of freedom is given, K is the stiffness matrix, and M is the mass matrix. The set is represented as a block matrix based on the principal and dependent degrees of freedom. For the set of accelerations from the degrees of freedom, The set of accelerations for the main degrees of freedom, f B The set of excitations for the main degrees of freedom. If the inertial forces at the docking interface are neglected, the structure undergoes a Guyan reduction transformation.

[0020]

[0021] Among them, [T i ] is the transformation matrix, For the set of displacements of degrees of freedom, Main Degrees of Freedom Displacement Set

[0022]

[0023] I is the identity matrix, Φ c Let be the mode matrix of the intrinsic modes.

[0024] Stiffness matrix of the reduced substructure and quality matrix They are respectively

[0025]

[0026]

[0027] K i M is the stiffness matrix of the complete model. i It is the complete model quality matrix; The stiffness matrix of the substructure. For the substructure quality matrix, [T i ] is the transformation matrix.

[0028] The equations of motion for the substructure after Guyan reduction are as follows:

[0029]

[0030] Right now, The stiffness matrix of the substructure. The substructure quality matrix, The substructure displacement matrix, This is the acceleration matrix of the substructure.

[0031] Guyan reduction theory is used to reduce the degrees of freedom of the complete simulation model to generate a reduced substructure model. Equations (1) and (6) show that the number and quality of the master degrees of freedom (master degree-of-freedom nodes) determine the error between the reduced substructure model and the original complete model, and will significantly affect the mass and stiffness matrices of the generated reduced substructure. The selection of master degrees of freedom for the finite element model of welded structures mainly follows these principles: (1) The master degrees of freedom should be symmetrical about the axis. (2) The positions where the model connects to other structures should be selected as master degrees of freedom. (3) They should be distributed as evenly as possible across the model.

[0032] 1-3: Construct a dynamic substructure model.

[0033] To assess whether the reduced sub-model can reflect the dynamic characteristics of the original model, this invention proposes a mode matching method. The differential equations of motion for the free modes of the reduced sub-structure are as follows:

[0034]

[0035] The expression for the displacement vector is

[0036]

[0037] Where, {φ} i Let ω be the amplitude of the i-th mode. i It is the angular frequency of the i-th mode. Substituting equation (8) into equation (7) yields...

[0038]

[0039] The modal frequency ω of the i-th order mode can be obtained using equation (9). iand mode shape vector {φ} i .

[0040] Modal analysis was performed on the complete model and the reduced substructure model, and the calculation results were compared. When the average error of all modes of the reduced substructure model within the effective mass coefficient of 0.9 is no more than 2% compared with the original model, and the mode shapes are consistent, the reduced substructure model is considered to match the dynamic characteristics of the original model, and its name is defined as dynamic submodel.

[0041] 1-4: Constructing the weld substructure model:

[0042] A quasi-static analysis with a unit load is applied to the complete model using the traditional structural stress method to identify the critical weak welds with low fatigue life. These critical weak welds, along with other welds of interest, form a set of welds of interest requiring evaluation and calculation. The weld toe plate thickness section nodes of the solid element set of welds of interest, or the weld toe nodes of the shell element set, are added as new master degree-of-freedom nodes to the master degree-of-freedom nodes of the dynamic sub-model, forming a weld substructure model.

[0043] Step 2 is as follows:

[0044] 2-1 Dynamic simulation of rigid-flexible coupling in dynamic substructure model;

[0045] To account for the deformation and vibration characteristics of the welded structure, a rigid-flexible coupled dynamic model is established using a dynamic substructure model to obtain the time-domain excitation of the vibration characteristics of the welded structure.

[0046] By incorporating constraints into the system equations using the Lagrange multiplier method, and then combining them with the equations of a multi-rigid-body system, the dynamic equations of a rigid-flexible coupled system can be obtained.

[0047]

[0048] In the formula, ψ(ξ,t)=0 represents the whole constraint; ξ represents the generalized coordinates of the flexible body, including the coordinate displacement X, Euler angle coordinates Ω, and modal coordinates q. * That is, ξ=[X Ω q * ] T Q represents the generalized force, including physical force; L is the Lagrange function, L = TV, which is the difference between kinetic and potential energy; λ is an undetermined factor; F is the dissipation function.

[0049] 2-2 Extracting the time-domain load generated from the dynamic simulation;

[0050] Step 3 is as follows:

[0051] 3-1: Transient analysis using the weld substructure model:

[0052] Transient analysis was performed using a weld substructure model to obtain the stress history of the weld's response structure. The equations of motion for the weld substructure are as follows:

[0053]

[0054] In the formula: It is the damping matrix of the weld substructure; It is the dynamic load history. It is the substructure velocity matrix.

[0055] Over a certain time interval, assume the integrals of the final velocity and displacement are:

[0056]

[0057]

[0058] In the formula: α and δ are integration constants; Δt is the time interval between the nth step and the (n+1)th step; from the above two formulas, we can deduce that the substructure at t n+1 acceleration at time and speed The matrix expression can then be used to obtain the substructure at time t. n+1 displacement at time speed and acceleration

[0059] {F(t)}=[K e ]{U(t)} (14)

[0060] According to equation (14), the nodal force matrix {F(t)} of the weld substructure element node response as a function of time can be solved, where [K e ] represents the stiffness matrix of the weld substructure element.

[0061] Step 4 is as follows:

[0062] 4-1: Obtaining the equivalent structural stress in the transient response:

[0063] The transient response structural stress at the weld toe line penetrating the weld plate thickness neutral plane (hereinafter referred to as the plate thickness mid-plane) is obtained, and the structural stress σ on the plate thickness mid-plane is obtained according to equations (15) and (16). s (t) and shear stress τ s (t) Transient response time domain value:

[0064]

[0065]

[0066] In the above formula, σ s (t), σ m (t), σ b (t), Fy (t) represents the transient response time history of structural stress, membrane stress, bending stress, and nodal force in the direction perpendicular to the weld toe line on the mid-surface of the plate thickness, respectively. x (t) is the transient response time history of bending moment along the weld toe line on the thick surface of the weld plate, τ s (t), τ m (t), τ b (t), F x (t) represent the transient response time histories of shear structural stress, shear film stress, shear bending stress, and shear nodal force along the weld toe line on the mid-thickness surface of the plate, respectively. y (t) is the transient response time history of bending moment in the direction perpendicular to the weld toe on the thick mid-surface of the weld plate, d is the plate thickness, and L is the distance matrix between nodes on the mid-surface of the weld thickness section in the structural stress method.

[0067] 4-2: Obtain the equivalent structural stress of the transient response at the mid-thickness surface of the plate.

[0068] Define the transient response load ratio correction factor:

[0069]

[0070]

[0071]

[0072]

[0073] Based on (2) to (7), the equivalent structural stress transient response values ​​are obtained:

[0074]

[0075]

[0076] In the above formula, S σ (t) is the transient response equivalent structural stress perpendicular to the weld toe line on the neutral plane of the weld thickness, S τ (t) is the transient response equivalent structural stress along the weld toe line on the thick surface of the weld plate, where m is a constant of 3.6;

[0077] When S τ (t)>S σ (t) / 3 hours

[0078]

[0079] otherwise

[0080] S e (t)=S σ (t) (24)

[0081] β is a constant representing the ratio of fatigue strength between the normal stress based on fatigue testing and the shear stress based on testing.

[0082] S σ (t) is the time-domain equivalent structural stress of the nodal response perpendicular to the weld toe line;

[0083] S τ (t) is the time-domain shear equivalent structural stress of the nodal response along the weld toe line;

[0084] S e (t) represents the time-domain equivalent structural stress of the nodal transient response.

[0085] 4-3: The time history of thermal equivalent structural stress is compiled into a thermal equivalent structural stress load spectrum of the time domain response of each node using the rainflow counting method, including the equivalent structural stress range ΔS of the i-th order (i = 1, 2…k). si and the number of loops n i ;

[0086] Substituting the load spectrum into the calculation formula yields the fatigue life cycle count at failure:

[0087] N i =(△S) si / C d ) 1 / h (25)

[0088] In the formula: C d and h are experimental constants for the structural stress method. N i For △S si The number of fatigue life cycles of the lower welded joint;

[0089] The fatigue life of the weld can be predicted, i.e., the cumulative fatigue damage ratio.

[0090]

[0091] As can be seen from equations (6), (7), and (12), by reducing the number of nodes in the main degrees of freedom of the substructure model, the dynamic characteristics of the complete model can be better reflected. The dynamic and transient response results can be obtained by calculating only these nodes, thereby reducing the degrees of freedom in the structural finite element analysis and achieving the goal of greatly improving the computational efficiency.

[0092] The fatigue life assessment method for welded structures based on transient structural stress of substructures in this invention, compared with traditional transient analysis methods, takes into account both the rigid-flexible coupling dynamic simulation of welded structures and the vibration characteristics of the welded structures themselves, and greatly improves the efficiency of transient analysis. It has important theoretical research and practical significance for the fatigue resistance design of welded structures. Attached Figure Description

[0093] Figure 1 Flowchart of this invention.

[0094] Figure 2 A simplified substructure model of the welded car body A600 for high-speed trains.

[0095] Figure 3 A schematic diagram of the weld seams in the aluminum alloy welded vehicle body.

[0096] Figure 4 Comparison chart of modal analysis results.

[0097] Figure 5 Comparison chart of modal analysis time consumption.

[0098] Figure 6 Comparison chart of transient simulation analysis time consumption.

[0099] Figure 7(a) Comparison of equivalent structural stress of weld 1 at a certain moment.

[0100] Figure 7(b) shows the comparison of equivalent structural stresses of weld 6 at a certain moment.

[0101] Figure 8(a) Comparison of fatigue damage in weld 1.

[0102] Figure 8(b) Comparison of fatigue damage in weld 6. Detailed Implementation

[0103] The present invention will now be further described with reference to an example of fatigue assessment of weld seams in the aluminum alloy welded car body of a high-speed train, and the accompanying drawings:

[0104] The finite element simulation computer hardware configuration is an Intel i7-8700 CPU 3.20GHz 8-core CPU and 32GB of RAM.

[0105] 1. Sub-model establishment:

[0106] First, a finite element model including the weld seam is established, with 814,687 elements and 637,328 nodes, named AFull model. Taking 600, 520, and 450 master degrees of freedom nodes as examples, reduced substructure models named A600, A520, and A450 are established respectively, condensing all finite element elements into a single superelement, such as... Figure 2 It is the A600 model.

[0107] Modal analysis was performed on the complete vehicle body model and the reduced substructure model using the Black-Lanczos modal extraction method. The modal comparison results are shown in [Figure number missing]. Figure 4 As the number of master degrees of freedom increases, the modal analysis time will also increase, but the modal results will be closer to the complete model.

[0108] The first nine non-zero effective mass coefficients of A600, A520, and A450 are 0.9, 0.9, and 0.88, respectively. Their highest modal errors are 3.33%, 10.82%, and 6.9%, respectively, and their lowest modal errors are 0.23%, 0.53%, and 0.79%, respectively. The average modal errors are 1.39%, 3.63%, and 3.54%, respectively. In other words, the lower the average modal error between the reduced sub-model and the complete model, the better it reflects the dynamic characteristics of the original structure. A600 has an effective mass coefficient of 0.9, an average modal error of less than 2%, and consistent vibration modes, meeting the modal matching requirements, and is therefore designated as the dynamic substructure model D600.

[0109] pass Figure 5 It can be seen that when applying substructure technology to perform modal analysis of high-speed train bodies, the average computation time of the substructure model is reduced to only 47.23% of that of the complete model, and the time consumed by modal analysis is significantly reduced.

[0110] 2. Dynamic simulation of rigid-flexible coupling:

[0111] The dynamic substructure model is imported into dynamics software (such as Simpack) for flexible processing. The vehicle body is regarded as a flexible body and connected to the rigid bogie through hinges. The suspension is simulated by force elements to obtain a rigid-flexible coupling model of the whole vehicle.

[0112] Using the German Level 5 track spectrum as the track excitation and a vehicle speed of 300 km / h as an example, a dynamic simulation of the rigid-flexible coupled model of the whole vehicle was performed. Because the high-speed train experiences a large vertical load, the vertical load of the air springs was simulated. The simulation time was 20 seconds, and the sampling frequency was 100 Hz. After the simulation, the vertical time loads at the four air springs, based on time-domain integration, were obtained as the external loads for transient simulation analysis.

[0113] 3. Transient simulation analysis:

[0114] For ease of comparison later, weld substructure models W600, W520, and W450 were generated based on the A600, A520, and A450 models according to the method in step 1.4. Seven weld locations of interest are listed below. Figure 3 As shown.

[0115] The time-domain excitation was used to analyze the time load history at the air spring. Finite element software (such as Ansys) was used to perform transient analyses on the complete vehicle body model and the weld substructure model using both the full method and modal method. For ease of comparison of calculation times, the simulation analysis duration was set to 20 seconds, the time interval to 0.1 seconds, and the number of welds of interest was increased to 10, 50, and 100, with an average of 30 elements per weld.

[0116] Figure 6Table 1 compares the time consumption results of transient simulation analysis. The modal superposition method has an effective quality coefficient of 0.99 and takes 70% of the time of the complete model. The dynamic substructure model is next, taking 41.9% of the time of the complete model because it needs to calculate the master and slave degrees of freedom. The method proposed in this invention takes the least time, only 2.26% of the time of the complete model. Furthermore, the time increases slightly with the number of welds of interest, and the time consumption of the weld substructure model with 100 weak welds is 670s, which is only 5.03% of the time of the complete model.

[0117] 4. Vehicle body fatigue life assessment:

[0118] According to the method in step 1.5, the fatigue life of the welds of interest is evaluated using a self-compiled software based on the structural stress method. According to the fatigue damage ranking, welds weld_1 and weld_6 have the lowest fatigue lives among the seven welds of interest. The equivalent structural stress and total fatigue damage at a certain moment in the transient response are shown in Figures 7(a), 7(b), 8(a), and 8(b). According to Tables 2 and 3, the substructure model proposed in this invention and the complete model show almost identical prediction results for the transient fatigue life of the vehicle body welds, with a maximum error of only 2.01% and a minimum error of only 0.5%. It is evident that the method of this invention significantly reduces the transient simulation calculation time while ensuring the accuracy of the results.

[0119] Table 1

[0120]

[0121] Table 2

[0122]

[0123] Table 3

[0124]

[0125] This invention proposes a fatigue life assessment method for welded structures based on transient structural stress in substructures, which significantly reduces the simulation time cost of transient weld fatigue assessment methods in engineering and makes them easier to use. Through this invention:

[0126] (1) Based on this method, the simulation time cost for fatigue life assessment of transient welded structures is significantly reduced, while the error of the life result is very small. Furthermore, welds can be flexibly added for new assessments, meeting the needs of practical engineering. Taking the fatigue life assessment of aluminum alloy welded car bodies of high-speed trains as an example, the simulation time was reduced from 3 hours to 11 minutes, the calculation time was reduced by 95%, the equivalent structural stress error was 1.1%, and the fatigue damage ratio error was 2%.

[0127] (2) The sub-model matching method effectively matches the dynamic characteristics of the complete model and reduces the time required for modal analysis. Examples show that when the sub-model has an effective mass coefficient of less than 0.9, the average error between all modes and the original model modes does not exceed 2%, and the mode shapes are consistent, the sub-model can match the dynamic characteristics of the original complete model well. The modal analysis time is 47.23% of that of the original complete model, and the weld fatigue assessment error based on this sub-structure model is very small.

[0128] (3) The simulation of rigid-flexible coupling of welded structures and transient fatigue simulation analysis were considered. Unlike the traditional quasi-static simulation evaluation, the fatigue evaluation was carried out by using the equivalent structural stress that takes into account the vibration characteristics of the structure itself, making the simulation more in line with engineering reality.

[0129] This invention reduces the cost of repeated prototype testing and significantly saves simulation time. The method proposed in this invention can be extended to other mechanical welded structural components, providing new ideas and methods for the dynamic fatigue resistance design of welded structures.

[0130] This invention has been described through embodiments. Those skilled in the art will understand that various changes or equivalent substitutions can be made to these features and embodiments without departing from the spirit and scope of the invention. Furthermore, under the teachings of this invention, these features and embodiments can be modified to adapt to specific situations and materials without departing from the spirit and scope of the invention. Therefore, this invention is not limited to the specific embodiments disclosed herein, and all embodiments falling within the scope of the claims of this application are within the protection scope of this invention.

Claims

1. A method for assessing the fatigue life of welded structures based on transient structural stress in substructures, characterized in that: The steps include: Step 1: Construct the reduced substructure model, dynamic substructure model, and weld substructure model, as detailed below: 1-1: The geometric model of the welded structure is meshed using the finite element method, with the mesh consisting of 2D or 3D elements; 1-2: Constructing a reduced substructure model: Choosing the interface degree of freedom as the principal coordinate, denoted by B; and the internal degrees of freedom to be reduced as the secondary coordinate, denoted by I, then for a multi-degree-of-freedom system neglecting damping, its equation of motion during free vibration is: in, For the set of displacements of degrees of freedom, The set of displacements for the principal degrees of freedom is given, K is the stiffness matrix, and M is the mass matrix. The set is represented as a block matrix based on the principal and dependent degrees of freedom. For the set of accelerations from the degrees of freedom, The set of accelerations for the main degrees of freedom, f B Given the set of excitations for the main degrees of freedom, and neglecting the inertial forces at the docking interface, the structure undergoes a Guyan reduction transformation. Among them, [T i ] is the transformation matrix, For the set of displacements of degrees of freedom, Main Degrees of Freedom Displacement Set I is the identity matrix, Φ c The mode matrix represents the intrinsic modes; Stiffness matrix of the reduced substructure and quality matrix They are respectively K i M is the stiffness matrix of the complete model. i It is the complete model quality matrix; The stiffness matrix of the substructure. For the substructure quality matrix, [T i [ ] represents the transformation matrix; The equations of motion for the substructure after Guyan reduction are as follows: Right now, The stiffness matrix of the substructure. The substructure quality matrix, The substructure displacement matrix, The acceleration matrix of the substructure; 1-3: Constructing a dynamic substructure model: The differential equations of motion for the free modes of the reduced substructure are as follows: The expression for the displacement vector is Where, {φ} i Let ω be the amplitude of the i-th mode. i It is the angular frequency of the i-th mode. Substituting equation (8) into equation (7) yields... The modal frequency ω of the i-th order mode can be obtained using equation (9). i and mode shape vector {φ} i Modal analysis was performed on the complete model and the reduced substructure model, and the calculation results were compared. 1-4: Constructing the weld substructure model: The traditional structural stress method is used to apply a unit load quasi-static analysis to the complete model to obtain the key weak welds with low fatigue life. The key weak welds and other welds of interest are combined into a set of welds of interest that need to be evaluated and calculated. The weld toe plate thickness section node of the solid element or the weld toe node of the shell element of the set of welds of interest are added as new master degree of freedom nodes to the master degree of freedom nodes of the dynamic submodel to form the weld substructure model. Step 2: Dynamic simulation of rigid-flexible coupling, as follows: 2-1: Dynamic simulation of rigid-flexible coupling in dynamic substructure model: A rigid-flexible coupled dynamic model is established using a dynamic substructure model to obtain the time-domain excitation of the vibration characteristics of the welded structure. Constraints are introduced into the system equations using the Lagrange multiplier method. The dynamic equations of the rigid-flexible coupled system are obtained by combining the equations of the multi-rigid-body system with those of the multi-rigid-body system. In the formula, ψ(ξ,t)=0 represents the whole constraint; ξ represents the generalized coordinates of the flexible body, including the coordinate displacement X, Euler angle coordinates Ω, and modal coordinates q. * That is, ξ=[X Ω q * ] T Q represents the generalized force, including physical force; L is the Lagrange function, L = TV, which is the difference between kinetic and potential energy; λ is an undetermined factor; F is the dissipation function. 2-2 Extracting the time-domain load generated from the dynamic simulation; Step 3: Perform transient analysis using the weld substructure model, as detailed below: 3-1: Transient analysis using the weld substructure model: Transient analysis was performed using a weld substructure model to obtain the stress history of the weld response structure and the motion equations of the weld substructure. In the formula: It is the damping matrix of the weld substructure; F i B (t) represents the dynamic load history. It is the substructure velocity matrix; Over a certain time interval, assume the integrals of the final velocity and displacement are: In the formula: α and δ are integration constants; Δt is the time interval between the nth step and the (n+1)th step; from the above two formulas, we can deduce that the substructure at t n+1 acceleration at time and speed The matrix expression can then be used to obtain the substructure at time t. n+1 displacement at time speed and acceleration {F(t)}=[K e ]{U(t)} (14) According to equation (14), the nodal force matrix {F(t)} of the weld substructure element node response as a function of time can be solved, where [K e [ ] represents the stiffness matrix of the weld substructure element; Step 4: Use the equivalent structural stress fatigue life assessment based on the transient response of the welded structure, as detailed below: 4-1: Obtaining the equivalent structural stress in the transient response: The transient response structural stress at the weld toe line penetrating the weld plate thickness neutral plane is obtained, and the structural stress σ on the plate thickness neutral plane is obtained according to equations (15) and (16). s (t) and shear stress τ s (t) Transient response time domain value: In the above formula, σ s (t), σ m (t), σ b (t), F y (t) represents the transient response time history of structural stress, membrane stress, bending stress, and nodal force in the direction perpendicular to the weld toe line on the mid-surface of the plate thickness, respectively. x (t) is the transient response time history of bending moment along the weld toe line on the thick surface of the weld plate, τ s (t), τ m (t), τ b (t), F x (t) represent the transient response time histories of shear structural stress, shear film stress, shear bending stress, and shear nodal force along the weld toe line on the mid-thickness surface of the plate, respectively. y (t) is the transient response time history of bending moment in the direction perpendicular to the weld toe on the thick mid-surface of the weld plate, d is the plate thickness, and L is the distance matrix between nodes on the mid-surface of the weld thickness section in the structural stress method. 4-2: Obtain the equivalent structural stress of the transient response at the mid-thickness surface of the plate. Define the transient response load ratio correction factor: Based on (2) to (7), the equivalent structural stress transient response values ​​are obtained: In the above formula, S σ (t) is the transient response equivalent structural stress perpendicular to the weld toe line on the neutral plane of the weld thickness, S τ (t) is the transient response equivalent structural stress along the weld toe line on the thick surface of the weld plate, where m is a constant of 3.6; When S τ (t)>S σ (t) / 3 hours otherwise S e (t)=S σ (t) (24) β is a constant representing the ratio of fatigue strength between the normal stress based on fatigue testing and the shear stress based on testing. S σ (t) is the time-domain equivalent structural stress of the nodal response perpendicular to the weld toe line; S τ (t) is the time-domain shear equivalent structural stress of the nodal response along the weld toe line; S e (t) represents the time-domain equivalent structural stress of the nodal transient response; 4-3: The time history of thermal equivalent structural stress is compiled into a thermal equivalent structural stress load spectrum of the time domain response of each node using the rainflow counting method, including the equivalent structural stress range ΔS of the i-th order (i = 1, 2…k). si and the number of loops n i ; Substituting the load spectrum into the calculation formula yields the fatigue life cycle count at failure: N i =(△S si / C d ) 1 / h (25) In the formula: C d and h are the experimental constants of the structural stress method, N i For △S si The number of fatigue life cycles of the lower welded joint; The fatigue life of the weld can be predicted, i.e., the cumulative fatigue damage ratio.

2. The fatigue life assessment method for welded structures based on transient structural stress of substructures according to claim 1, characterized in that: In steps 1-4, when the average error between all modes of the reduced sub-model and the original model modes is no more than 2% within an effective mass coefficient of 0.9, and the mode shapes are consistent, the reduced sub-structure model is considered to match the dynamic characteristics of the original model, and its name is defined as dynamic sub-model.

Images